The overall goal is to obtain data needed to prevent debilitating outcomes of child and adolescent psychopathology. This goal will be achieved (a) by identifying predictors of good versus poor outcomes following clinical referral, and (b) by making in-depth clinical comparisons of the best and worst outcomes cases. Outcomes will be assessed using multisource data for 2,461 American and 2,285 Dutch youngsters who were uniformly assessed at intake into mental health clinics. Specific aims are to: (1) test the ability of Child Behavior Profile types to predict outcomes assessed by multiple data sources; (2) determine the contribution of child and family characteristics, diagnoses, services, and other variables to the prediction of outcomes; (3) test the ability of Child Behavior Profiles and other initial variables to predict service dosage; (4) cross-validate predictive relations for outcomes and service usage in the U.S. and Holland (5) to direct clinical evaluation and PSM diagnoses of the best and worst outcome bases (6) In concordance among outcome data from multiple sources and construct cross-validated operational definitions of outcome that optimize relations to predictor variables. The predictive findings and clinical outcome data will be used to design interventions targeted on the needs of the poor outcome cases. Such interventions would then be tested in future research. In Phase 1 of the study, parents who completed standardized rating forms at intake will report on their child's current functioning, changes since intake, new problems, other services obtained, and current need for help. Teachers will rate academic performance, adaptive functioning, and behavioral/emotional problems. Subjects aged 12-27 will complete self- reports. Public access documents will be used to identify criminal convictions and suicides. In Phase 2, 100 American subjects having the best and 300 having the worst outcomes will be clinically interviewed, as will parents of those aged 6- 18. Comparisons of the best and worst outcome cases will provide a basis for designing better ways of helping those who currently have poor outcomes.